Oil Forecasters Say Hold off on Buying That Hummer

 

Takeaway

  • Oil’s likely to head upwards, perhaps to $100 a barrel in five years.
  • Now is a good time to acquire oil and gas reserves, while prices are low.
  • New statistical tool helps small energy firms evaluate long-term investments.

Don’t go buying that Hummer just yet. That’s the advice to America’s drivers from two researchers at the McCombs School of Business.       

Sure, a barrel of crude is around $50, and gasoline prices are at levels last seen in the depths of the Great Recession. But such prices aren’t likely to last, warns Jim Dyer, professor of information, risk and operations management, and clinical associate finance professor Joe Hahn.

Strip away temporary market fluctuations, they say, and the underlying price of a barrel of oil today is closer to $80. Five years from now, it could be flirting with $100 again.

“There’s already a big slowdown in drilling because lower prices have caused oil companies to cut back their capital investment programs,” says Hahn. “Market forces say we will stop growing oil production as much as we have been for the past few years, and because of that, prices will go back up.”

They can tease out long-term price trends from short-term ups and downs because they’ve seen it all before: on their computers. Their study, published in Energy Economics, analyzes 23 years of oil prices to model what lies ahead.

Their goal is not to spread gloom and doom but to offer a new tool to oil and gas producers. Those firms have to make investments today while guessing at prices 10 years out. “My interest is driven by practical considerations,” says Dyer. “Energy prices are a big concern for energy companies and how they make investment decisions.”

The two professors work with statistics now, but both have energy backgrounds, which might explain why their tool is a hybrid of two models. One is typically used for valuing assets in the financial markets, while the other models the fundamentals of oil supply and demand.

Both simulate what mathematicians call stochastic processes. On the surface, data look random, but underneath, patterns can be found. The ebb and flow of traffic congestion or shoppers in a checkout line can be modeled in similar ways.

For Dyer and Hahn, the first model estimates a long-term equilibrium price for oil. Like the Dow Jones or S&P indices, they expect that price to gradually rise over the long haul. The other model then projects how much prices will bounce up and down around that equilibrium in the short term.

Combining the two, they get a graph that looks like a tracking map for a hurricane. It shows the most likely path, surrounded above and below by a cone of uncertainty. That cone predicts how low or high prices might go at any given time. As you might expect, the uncertainty gets wider the farther ahead a forecast goes.

To reduce their own uncertainty, the forecasters calibrated their tool, checking its results against spot prices for West Texas Intermediate Crude from 1990 to this May. But to forecast the future, they deconstructed another set of data: prices of oil futures that were traded during the same 1990 to 2015 time period.


Oil futures markets, explains Dyer, are already looking ahead. “Those futures prices result from the people who are trading, thinking really hard, and using all the information that’s available to them at any one point in time, estimating what prices will be going forward.”

They acknowledge that there’s one big question their model can’t answer with certainty: To what degree has hydraulic fracturing fundamentally changed the dynamics of oil markets? If recent futures data underestimate that impact, prices could turn out to be lower than their model predicts. Says Dyer, “It’s hard to say at this point just how significantly the shift in technology with fracking will affect the long-term equilibrium price.”

He takes some comfort that today’s prices actually fit the lower range of his forecast. His cone of uncertainty bottoms out around $55 a barrel — which also indicates that prices are likely to trend upward in the future.

“Precise forecasting is very difficult to do, as proven over and over again through history,” says Hahn. “What companies are really interested in is that possible cone of uncertainty.”

If his projections are right, they have implications for today’s energy investors. In spite of all the drilling rigs sitting idle, it’s a good time to invest in drilling rights, because prices are likely to be higher five or 10 years from now.


“Now would be a good time to make acquisitions,” adds Hahn. “It’s a good time if you have cash because those reserves are likely to be worth a lot more in the future.”

Some companies are already taking the hint. In April, Royal Dutch Shell agreed to spend $70 billion for natural gas producer BG Group. The deal will boost Shell’s reserves at least 25 percent in locales like Brazil, East Africa, and Australia.

The professors hope their tool will be practical for smaller firms that need to make similar decisions but can’t invest in their own supercomputers.

He adds that because his tool relies on a single source of information – futures prices – he expects such firms will supplement it with information from other sources. “Any prudent company would look at this information,” he says, “but look at an up-to-date fundamentals forecast, too, and use their judgment and intuition.”

 

Footnote, figure 1: This depicts historical West Texas Intermediate spot prices (to May 12, 2015), followed by a forecast developed from a model of the oil price stochastic process.  To show historical forecast performance, the figure also includes overlays for forecasts that would have commenced in January 1996 and January 2005.  In all cases, the oil price process was specified using historical market futures prices dating from January 1990 to the start date of the forecast, as described in Hahn, DiLellio and Dyer (2014). A stochastic process model allows development of expected future prices, indicated by the bold solid lines in the figure, as well as a confidence envelops (the lighter lines showing the 10th and 90th percentiles in the figure) around the expected future values.  The latter depicts the uncertainty in the forecast, and it naturally increases in width as the length of forecast time increases. These confidence intervals also show how significantly the uncertainty associated with the most recent forecasts has increased, indicating increased volatility of the oil price process.

 

Footnote, figure 2: This depicts historical West Texas Intermediate spot prices (to May 12, 2015), followed by a forecast developed from a model of the oil price stochastic process.  The oil price process was specified using historical market futures prices dating back to 1990, as described in Hahn, DiLellio and Dyer (2014).  A stochastic process model allows development of expected future prices, indicated by the bold solid line in the figure, as well as a confidence envelop (the lighter lines showing the 10th and 90th percentiles in the figure) around the expected future values.  The latter depicts the uncertainty in the forecast, and it naturally increases in width as the length of forecast time increases. 
 

Related articles

Faculty in this Article

Jim Dyer

Professor McCombs School of Business

Jim Dyer is a professor in the department of Information, Risk, and Operations Management. He received his B.A. and Ph.D. from The University...

Joe Hahn

Clinical Associate Professor McCombs School of Business

Joe Hahn received his B.S. and M.S. in engineering from the University of Texas at Austin followed by an MBA and Ph.D. in management science and...

About The Author

Steve Brooks

In a quarter-century as a journalist, Steve Brooks has won two Neal awards for excellence in trade reporting and a Press Club of New Orleans award...

Leave a comment

We want to hear from you! To keep discussions on-topic and constructive, comments are moderated for relevance and for abusive or profane language.
Login or register to post comments